Adverbs as evidentials : an English-Spanish contrastive analysis of twelve adverbs in spoken and newspaper discourse

This paper presents a contrastive analysis of six English evidential adverbs ending in -ly with their Spanish nearest translation equivalents, in spoken and newspaper discourse. The adverbs may be associated with varying degrees of reliability: high (clearly/ claramente, evidently/evidentemente, obviously/obviamente), medium (apparently/al parecer) and low (seemingly/aparentemente, supposedly/supuestamente). The analysis is based on tokens of authentic language extracted from two contemporary corpora, the Corpus of Contemporary American English (COCA) and the Corpus de Referencia del Español Actual (CREA). The qualitative analysis focuses on the evidential functions of the adverbs and on their pragmatic interactional uses; the quantitative analysis centres on the relative frequency of type of evidential functions and the clausal position of the adverbs. The results uncover a number of differences between the English adverbs and their Spanish correlates and also between the two discourse types. Practically all the adverbs are strongly specialized in expressing either indirect-inferential or indirect-reportative evidentiality. English obviously and Spanish evidentemente show a high frequency of cases of loss of evidential meaning due to pragmaticalization, specifically in spoken discourse. Regarding position, the English adverbs are more frequent in medial clausal position, while some Spanish adverbs are often found in the more prominent parenthetical position.


Introduction
Evidentiality concerns the source of information or evidence a speaker/writer invokes as justification for making a claim (Anderson 1986;Boye 2012).Justification may make reference to direct access to the evidence through perceptual sources, as well as indirect access, through inference or mediated through report or hearsay (Chafe & Nichols 1986;Willett 1988;Plungian 2001;Aikhenvald 2004;Boye 2012, inter alia).
Within the field many scholars have restricted the notion of evidentiality to cases of grammatical markers in a narrow sense (cf.Anderson 1986).Aikhenvald (2007, 222) has argued that "the term 'evidential' is best used for closed grammatical systems, and the term 'information source' for the vast body of other ways of referring to knowing things".However, Wiemer (2010, 60) proposes a wider scope, from a functionalonomasiological perspective, including "all kinds of distinct means serving to express evidential functions, especially if they convey some specific meanings from a taxonomy of evidential functions", on condition that these linguistic devices or function words "can be regarded as sufficiently conventionalized (and not only as an evidential strategy), i.e., with a stable, non-detachable indication" of an evidential function.In Germanic and Romance languages, like English or Spanish, which do not have obligatory grammatical evidentiality, the conceptual domain of evidentiality is configured by some expressions from the closed subclasses of modal verbs, parentheticals, particles, etc., together with a number of adverbs, adjectives and lexical verbs (cf. Lampert & Lampert 2010;Wiemer 2010;Marín-Arrese 2015).Evidentials of the second type have traditionally been considered as lexical devices, due to the word class to which they belong.However, evidential adverbs are grammatical according to all the criteria proposed in Boye and Harder's (2012) usage approach to grammaticality and grammaticalization.The adverbs fulfil the non-focalizability criteria (Boye & Harder 2012, 14): they cannot occur in the focal position of cleft or pseudo-cleft constructions nor in the semantic scope of focus particles such as only, just, and even, nor can they receive focal stress. 1 They also fulfil the non-addressability criteria (Boye & Harder 2012, 15), since they cannot independently be questioned by WH-questions or yes-no questions, nor referred to anaphorically or cataphorically.
Evidential adverbs in English and European languages are a fruitful research area, as attested by many studies (Simon-Vandenbergen & Aijmer 2007;Celle 2009;Cornillie 2010;Carretero & Zamorano-Mansilla 2013;Lavid et al. 2016Lavid et al. , 2017;;Ruskan 2015;1 The adverbs can receive contrastive stress, but this possibility is not a criterion against grammaticality since it is shared by other clearly grammatical devices such as tense or aspect.Wiemer & Socka 2017).This paper sets forth a contrastive study of six English adverbs in -ly and their Spanish nearest translation equivalents, all of which can express evidentiality but also have non-evidential readings.The study is based on occurrences extracted from two corpora of contemporary language, the Corpus of Contemporary American English (COCA) and the Spanish Corpus de Referencia del Español Actual (CREA).The registers chosen are spoken and newspaper discourse, which differ in mode, spontaneity and degree of planning.The selected data consist of 100 examples of each adverb, 50 from spoken discourse and 50 from newspaper discourse.We have thus opted for a relatively large number of adverbs and for the analysis of a small number of cases of each, so that the resulting study is tentative but, on the other hand, has the advantage of being comprehensive.
The qualitative analysis focuses on the evidential functions of the adverbs and on their pragmatic interactional uses.The quantitative analysis aims at a cross-language and cross-register comparison in terms of three dimensions: a) evidential and non-evidential functions; b) relative frequency of different subtypes of evidentiality; c) clausal position of the evidential occurrences.Overall frequencies and percentages are discussed, but no inferential statistics are provided given the limited number of examples.
The paper is structured as follows: Section 2 describes the concept and types of evidential functions this research focuses on, the pragmatic extensions of evidential adverbs and the relation between evidentiary validity and evidentiary sources.Section 3 describes the corpora and specifies the data selected and the method designed for the quantitative and qualitative analysis.Section 4 provides a qualitative description of the adverbs, organized in pairs formed by an English adverb and its nearest Spanish equivalent.The results of the quantitative analysis are presented and discussed in Section 5. Finally, Section 6 sums up the main conclusions and suggests pointers for further research.

Evidentiality
Within the domain of evidentiality, we find various classifications of evidential values based on dimensions such as the type of evidence, mode of knowing and source of evidence (Chafe & Nichols 1986;Willett 1988;Plungian 2001;Squartini 2008;Cornillie et al. 2015;Marín-Arrese 2015).In this paper we focus on the following categories of evidentiality (cf.Diewald & Smirnova 2010; Marín Arrese 2015, 2017): (a) Direct Perceptual evidence (DPE): These are expressions indicating personal, direct access to visual or some other sensory form of evidence, external to the speaker/writer.In our corpora samples, we have only found the following example: (1)

Evidential Adverbs: Semantic Scope and Position
There seems to be an almost general consensus in the literature that evidential expressions have propositional scope.This is what usually distinguishes an evidential marker from an expression describing or designating some form of evidence.Apart from the basic condition that the meanings may be described in terms of the notion of 'evidence', Boye (2010, 304) argues that "for a given linguistic expression to be considered as having evidential meaning, it must be attested with a proposition-designating clause as its semantic scope", that is, not a state-of-affairs-designating clause.This paper focuses on evidential sentence adverbs, and on parentheticals, with an explicit propositional scope as in (4), and for the case study does not consider adverbs with an implicit propositional scope, as in the elliptical clause in (5), or adverbs having phrasal predicational scope such as (6), or as modifiers within a noun phrase in (7) (cf.Ruskan 2015): (4) In view of the results of previous research on evidential adverbs (Simon-Vandenbergen & Aijmer 2007; Ruskan 2015; Carretero to appear), we predict that many of the adverbs will occur frequently in more than one position.Nuyts (2001, 262-272) offers a functionalprocedural account for the positional freedom of epistemic modal expressions, including evidential adverbs such as presumably and supposedly, in terms of the interaction of two functional forces: an iconic force, which highlights the conceptual meta-status of the expressions over a proposition, and an information-structural force, which pushes the expression to non-focal positions.Nuyts (2001, Chapter 5) reports an experiment involving controlled data elicitation, described in more detail in Nuyts and Vonk (1999), according to which epistemic modal expressions have a tendency to be non-focal.This tendency is coherent with Boye and Harder's (2012, 14) non-focalizability criteria, mentioned in Section 1. Nuyts (2001) also observes that parenthetical uses and nonprominent positions (such as medial position) favour non-focality.We believe, however, that parenthetical position is iconic in the sense that intonational or orthographic separation from the rest of the clause represents its status as a meta-operator over the clause.As for initial position, we may consider that it gives weight to iconicity, since the evidential operator precedes the clause under its scope even if the adverbs do not provoke as strong an iconic effect.In its turn, final position will predictably be an odd choice for the adverbs under study, since it does not favour either of the two functional forces.It is not iconic, since it does not highlight status as meta-operator; as for information structure, this position is at odds with the tendency of the adverbs to be non-focal: in languages that tend to follow the principle of end-focus, as is the case in English and Spanish (Leech & Short 2007, 170-172;Lavid et al. 2010, 354-356) final position is a powerful resource for expressing focality.Due to the focalizability constraints for these adverbs (see Introduction), if placed in final position they do not seem to be focal but to lie outside the focal part of the clause.This non-focality may be seen in ( 14): ( Even if we can only guess, since the COCA transcription does not provide intonation features, focus sounds more natural in 'a big year' than in 'obviously'.The placement of the adverb in final position is perhaps due to the speaker's wish to downtone the focality of 'a big year' in order to sound less assertive for reasons of politeness, since the event communicated is one to which the addressee has privileged knowledge, i.e. a B-event in terms of Labov (1972, 254).
Therefore, in terms of the two forces, medial position is geared at not conferring prominence in discourse to the adverbs, while initial and parenthetical positions lay more emphasis on signalling its status as a meta-operator over the clause.The study of the relative frequency of each of these positions will then shed light on the extent to which language users give priority to one force or the other.Final position will predictably be uncommon, since it does not favour either of the forces.

Evidentiary Validity and Speaker Commitment
Another dimension of evidentiality was proposed by Chafe (1986, 262-263), who signals as one of its characterizing features the implications of the reliability of the information.Marín-Arrese (2013) has pointed out that a distinction needs to be made between the reliability (conventionally) assigned to the source of evidence and the degree of commitment of the speaker/writer with respect to the validity of the information.The different types of evidential expressions may carry some indication of speaker/writer attitude and commitment towards the validity of the communicated information.Davis et al. (2007, 73) have plausibly argued for the pragmatic fact that "some evidential morphemes are perceived to be stronger than others, and that this can, in turn, impact perceptions about the speaker's commitment to the main-clause content".
Within the literature there has been considerable discussion on the issue of the reliability of the sources of evidence.Traditionally, as noted by Willett (1988), observational evidence has been assigned a higher degree of reliability, whereas evidence based on reasoning suggests a lower degree of reliability of the source.Willett (1988, 57) proposes the following hierarchy of reliability of evidentiary sources: personal experience » direct (sensory) evidence » indirect evidence » hearsay Matlock (1989, 215) appears to correlate mode of access to the evidence with degree of speaker/writer certainty: "direct experience corresponds to a high degree of certainty, indirect experience of the reported type corresponds to less certainty, and indirect experience of the inferred type corresponds to even less certainty".However, the degree of strengthening of the assertive force of the evidentially modified utterance will depend on contextual factors and subjective probabilities, since direct perceptual evidence from an unreliable witness may carry less weight than the hearsay report of a highly trustworthy source.As Davis et al. (2007, 80-81) point out, "Though direct evidence might be reliably better than hearsay evidence, this is not a lexical fact per se, but rather a fact that we derive from general regularities in the world and the context of utterance, and thus it is conceivable that things could be reversed in some situations".Similarly, Marín-Arrese (2013, 422) argues that "The different modes of access to the information correlate with different values on a scale of speaker/writer's commitment towards the validity of the information.The various ways of framing a proposition (P) present a highly complex system of epistemic positioning strategies that speakers/writers have at their disposal for legitimising assertions".A case in point is that of reportative or mediated evidentiality, where, as Marín-Arrese (2013, 423) has observed, "the presumed validity of the information would rely to a great extent on the reliability of the original source of the information, whether s/he is considered an expert in the relevant field, or having a prestigious social status, or whether the information is considered warrantable on the basis of its widespread or universal acceptability".
Bearing these caveats in mind, in this paper we have grouped English clearly, evidently and obviously and their Spanish correlates as expressions indicating a source of information conventionally high in reliability, so that the communicated information is esteemed as high in validity.The adverbs apparently and al parecer might be considered to indicate medium validity of the information, since they allow the authorial voice to adopt a neutral position with regard to the reported information.The adverbs seemingly, aparentemente, supposedly and supuestamente seem to indicate that the authorial voice distances him/herself from the communicated information, thus indicating a lower degree of commitment regarding the validity of the information.According to the OED, seemingly may express a meaning of "?So as to seem real", and the Merriam Webster Dictionary defines the adjective seeming as "outwardly or superficially evident but not true or real".The OED also notes that one of the meanings of the adjective seeming is that of "a.apparent to the senses or to the mind, as distinct from what is".Similarly, for supposedly, the OED gives as one of its meanings that of "as a pretense, by way of feigning".As part of their meaning component, these adverbs seem to share the notions of 'illusory, feigned', which no doubt motivates their pragmatic use in the discourse, signalling non-alignment of the authorial voice with the external voice.
A final word in this section is in order regarding the pragmatic extensions of some of these evidential adverbs.Brinton (2005, 144) notes that "Modern English parentheticals such as I think/suppose/guess (subjective) or it seems (objective), <...> in addition to epistemic and evidential meaning, also serve purposes of intimacy and "positive" politeness (self-effacement and deference)".In the case of evidential adverbs, as Ruskan (2015, 104) has pointed out, traces of pragmaticalization tend to correlate with "high frequency, positional mobility (initial, medial, final) and scopal variability (clausal, phrasal)".When we observe the markedly higher frequencies of occurrence of some of these adverbs in our corpora of oral discourse in contrast with their use in journalistic discourse (see Table 1, in Section 3), we find that these criteria apply basically to high validity expressions, namely the adverbs obviously in English and evidentemente in Spanish.Our results for these adverbs agree with the description that their functions as pragmatic markers are to "establish a common ground with the addressee, emphasise the author's argumentation and link units of discourse (Brinton 2008)" (Ruskan 2015, 105).Similar statements are also found in other intralinguistic and crosslinguistic studies such as Simon-Vandenbergen and Aijmer (2007) or Cornillie (2010).In this paper, the pragmatic uses of the adverbs under study are discussed in the corresponding subsections.

Data Selection and Method of Analysis
This paper presents a case study on the evidential values of sentence adverbs, based on authentic examples from corpora of contemporary language.For English, the corpus selected is the Corpus of Contemporary American English (COCA), with over 520 million words, and more specifically the sections on spoken discourse and newspapers.For the Spanish corpus, Corpus de Referencia del Español Actual (CREA), which contains approximately 160 million words, we have also selected the sections on spoken discourse and newspapers.The sample used for the analysis consists of 100 randomly selected occurrences of each adverb, 50 from spoken discourse and 50 from newspaper discourse.The selection of examples was restricted to the occurrences of sentence adverbs with clausal scope (see 2.2); occurrences of adverbs with non-sentential scope were discarded.For two Spanish adverbs, obviamente and supuestamente, the total number of spoken occurrences in the CREA was smaller than 50, so the sample was enlarged with occurrences found in CORPES XXI, also part of the data bank of the Royal Academy of Spanish Language (Real Academia Española, RAE).
The examples selected were subjected to a qualitative and quantitative analysis.The quantitative data cover two dimensions: 1.The types of evidentiality proposed in Section 2.1, based on the mode of access to the information: Direct-Perceptual (DPE), Indirect-Inferential (IIE) and Indirect-Reportative (IRE).In the fourth category, Non-Evidential (NE), we grouped all the cases of sentence adverbs with non-evidential uses: non-evidential meanings such as appearance or manner, pragmaticalized uses, etc.
2. Position, which was registered for the evidential occurrences in terms of the subcategories specified in 2.2: initial (IN), medial (MD), final (FN) and parenthetical (PTH).
The description of the adverbs is organized in subsections, ordered from high to medium and low evidentiary validity, each covering an English adverb and its nearest Spanish correlate.Table 1 provides the basic information of number of words for the subcorpora and total number of tokens for each of the adverbs, as well as their normalized frequencies.

English and Spanish Evidential Adverbs
This section provides a qualitative account of the adverbs, organized in pairs formed by an English adverb and its nearest Spanish equivalent.For the description of the meaning of the adverbs, in certain cases we have cited the dictionary definitions provided by the Oxford English Dictionary (OED) or the Diccionario de la Lengua Española, Real Academia Española (DLE, RAE) for the sake of clarity.The quantitative results of the analysis are presented and discussed in Section 5.

Clearly and Claramente
Clearly and claramente have an evidential meaning ('it is clear that'), as in (15), and a non-evidential (NE) meaning of manner ('in a clear way'), as in ( 16).The CREA word count corresponds to the category 'press', including newspapers and magazines: the corpus website provides no individual count for each of these subcategories.However, all the analyzed examples have been extracted only from newspapers for the sake of comparability with the COCA.
3 Evidential clearly and claramente can marginally be graded by some adverbs such as very and its Spanish equivalent muy, but not with others such as more, most or almost and their Spanish equivalents.
(16) Lo que había sucedido en la asamblea de los consejos no se ha explicado demasiado claramente <NE> (CREA, 01.Industrias diversas.El Mundo, 2003).'What had happened in the assembly of the councils has not been explained too clearly.' The type of evidentiality is IIE in all the cases except for one direct-perceptual occurrence of clearly (Section 2, example 1).The strong reliability expressed by this adverb indicates that the inference is heavily based on evidence, which may be cognitive, as in example (15) above, or perceptual as in ( 17): (17) Someone in the room described the scene to us this way: We listened to it, and its clearly <IIE, MD> not Brian Kelleys voice.(COCA, SPOK: CBS_Sixty, 2003) The non-evidential occurrences typically express manner, except for a pragmaticized case of clearly (18), where the speaker/writer expresses insistence due to face-saving reasons: there is a concessive relationship between the statement qualified with clearly and the following face-threatening statement that the president should not apologize until the facts are clarified.
( The persuasive value is also strong in the three cases found of spoken clearly in clauses starting with and, where the speaker emphasizes that the clause in which the adverb occurs provides additional evidence to prove a fact, which accumulates with evidence provided in previous clauses, as for example, in ( 19): (19) The Whizzinator is a false penis to which is attached a, sort of a catheter and another source of urine.That's one of the things that the NFL football player, Onterrio Smith, was caught with.And it's clearly <IIE, MD> to avoid detection of doping substances in a urine test.(COCA, SPOK: NPR_FreshAir, 2006)

Evidently and Evidentemente
The meanings of evidently are defined by the OED as follows: "1.So as to be distinctly visible or perceptible; with perfect clearness, conspicuously <...>; 2. So that the fact predicated is evident; manifestly, obviously.<...> †3.Sc.Law.By evidence of a deed or document.Obs.".The first and the second meanings also hold for evidentemente.
The second meaning, which is found in both evidential and non-evidential cases, lays emphasis on the validity of the information.The evidential meaning occurs when this validity is qualified in terms of evidence, typically IIE.Depending on the individual cases, the inference may be perceptual-based (20), or conceptual-based ( 21).
( Due to this meaning of strong validity, the inference is presented as non-challengeable and (having the potential to become) part of the common ground between the speaker and the addressee.This feature favours their use in argumentative discourse as a resource for speakers/writers to persuade the addressee.
In non-evidential cases, the validity of the information is also assessed, but in terms of agreement with expectations or common sense rather than evidence, of course and naturally being adequate paraphrases, as in (22).Ruskan (2015, 114) reports a similar behaviour of the Lithuanian equivalent adverbial aišku 'clearly, of course', which "shows a bleaching of evidential functions and displays features of a pragmatic marker indicating common knowledge and interaction with the addressee".
(22) I'm not a lawyer, but evidently <NE> I asked for a read-out on this.He is actually, it was a legal maneuver to prove --it had nothing to say he was guilty.His lawyer told him to do it.But it's not him backing off on the confession.(COCA, SPOK: Fox, 2013) As may be seen in Section 3, Table 1, evidentemente is by far the most common of all the Spanish adverbs in spoken discourse, in sharp contrast to evidently, which is the least common in this discourse type.As was stated above, both adverbs easily present information as common ground.For this reason, they may have the interactional function of bonding, which consists in "the creation of shared attitudes, a common world" (Simon-Vandenbergen & Aijmer 2007, 154).For example, in (23), the first speaker expresses an inference drawn from the two accidents, 'I'm bad with doors', and the second speaker picks it up with 'Evidently', thus laying emphasis on common knowledge: (

Obviously and Obviamente
The meanings of obviously are defined by the OED as follows: "1.In a clearly perceptible manner, evidently, plainly, manifestly; naturally, as might be expected from the circumstances.<...> †2.By the way, in passing, incidentally.Obs." No evidential reportative values were found for obviously.Inferential evidential uses tend to draw on the notions of 'clearly perceptible' and 'manifestly', as illustrated in the following example: (24) RON-WARD-1FATHER# This is the, the x-ray photo that was sent to us.There's no tag numbers.MONA-WARD-1MOTHER# No. TIM-WILLIAMSON-1A# Okay.Cause obviously <IIE, PTH>, they've been tampered with.JIM-AVILA-1-ABC-N# (Voiceover) For years, Janie's parents for a new investigation to look into the witnesses story.(COCA, SPOK: ABC_Primetime Live, 2008) The conceptual-based inferential sense of obviously involving personal knowledge and/ or knowledge of the world is also quite frequent, as illustrated in the following example: (25) MAX-IRONS# Well, I thought great.But then I thought they have obviously <IIE, MD> made a mistake.They must have been drunk when they were casting.Because why I cast an English actor as an Austrian opera singer because I can't either speak Austrian or sing opera.(COCA, SPOK: NBC, 2015) The sense of 'naturally, as might be expected from the circumstances' is very often found with obviously.This meaning would not belong within the domain of evidentiality proper, since in most cases it does not involve any inference based on evidence, as in (26).We have thus categorized these cases as non-evidential uses.
(26) How hard is it to get up there?SAVIDGE: It is a problem.I mean 4 obviously <NE>, we go when invited and give inbeds, as they are called, by the U.S. military.

4
The combination 'I mean obviously' was found 72 times in the COCA Spoken corpus, and only once in the newspaper corpus.'American intelligence may have other information in this respect, as Bill Clinton suggested.We have evidence that there are obviously differences of opinion which are being formed in Belgrade, we have seen it this week, and there are things we know that I think we should comment.' 5 The co-occurence 'Obviously, I would' was found 29 times in the COCA Spoken corpus, and 4 times in the news corpus.
Non-evidential readings, lacking an inferential value, are quite frequent with obviamente, again with the sense of 'naturally', or with a confirmation value, as in ( 30 As in the case of other expressions indicating appearance (appear, seem), apparently has extended its meaning into the domain of reportative evidentiality (cf.Marín-Arrese 2017) and consistently shows the highest frequencies for this value in the different genres, as illustrated in ( 33): (33) When we made them clearer, it became apparent that certain things that were apparently <IRE, MD> on the tapes weren't there, and we gave our opinion with respect to those issues.<...> MR-ARMSTRONG: Well, the --the "n" word -nigger --was, uh, supposedly stated by Mr. Ulrich, and that is simply not there.Uh, we, uh --our experts say that.(COCA, SPOK: PBS_Newshour, 1996) From a pragmatic perspective, we find examples where a 'fake' inferential reading is exploited for purposes of irony and humour, as in ( 34).
(  Filomúsica, 11, 2000) 'Now that Karol has it all (he has even learnt French), all he needs is to recover Dominique.But that is not easy.He will only make her come with a stratagem: he appoints her as heir to all his goods, and then fakes his own death, buying a corpse with the face disfigured (apparently, with money one can buy almost anything), which is identified as Karol by all his family and acquaintances, who are in on it.' The conjunctive locution al parecer seems to have specialized for the reportative meaning, and also tends to appear as a parenthetical to a greater extent than apparently.
Crosslinguistically the parallelism between these two expressions of evidentiality is also present in the case of seemingly and aparentemente, where we find a certain balance between IIE and IRE meanings.

Seemingly and Aparentemente
In the OED we find the following meanings for seemingly: "2.a.To external appearance, apparently.<...> †b.?nonce-use.?So as to seem real.Obs.<...> 3. a.So far as it appears from the evidence; so far as one can judge by circumstances.<...> c. parenthetically.As it seems." Inferential evidential uses of seemingly appear to be based on the notions of both "as it appears from the evidence" and "so far as one can judge by circumstances", as illustrated in the following example: (37) # But the current public mood, laced as it is with racist attitudes, seemingly <IIE, MD> doesn't distinguish between legal and illegal immigrants.That means that even people who are obeying the law, working for a living, and paying taxes -the legal immigrants -are being targeted and penalized.# (COCA, NEWS: Denver Post, 1997) Reportative uses, as in (38) are also relatively frequent.In this respect, the meaning extension of seemingly appears to behave similarly to other evidential expressions invoking appearance.
( Non-evidential uses are also found invoking the meaning of external appearance, with the sense of 'real only in appearance', as in the following example: (39) # On balmy days, the doors are open, seemingly <NE> bringing the vineyards into the mostly white rooms, decorated with tapestry-like fabric in a muted grape and fruit motif, modern art and bouquets of fresh flowers.#... (COCA, NEWS: San Francisco Chronicle, 1990) According to the DLE, the adverb aparentemente has the meaning: "1.adv.Con apariencia."(with appearance).This meaning of 'manifest to the sight' or 'visible' is found in the following example of inferential, perceptual-based, evidentiality: (40) "Habían muerto hacía poco tiempo, aparentemente <IIE, IN> les habían robado, pues los bolsillos, con cierre de cremallera, estaban abiertos y vacíos", según Tawab.(CREA, PRENSA: La Razón, 2001).'They had died recently, apparently they had been robbed, since their pockets, with zippers, were open and empty", according to Tawab.' An inferential meaning based on the notion of 'judgement from circumstances' can also be found, as illustrated in the next example: (41) "La economía está, aparentemente <IIE, PTH>, en una senda de crecimiento muy por debajo de la tasa potencial realzada por la productividad, y aun después de las medidas que tomamos en enero, se mantiene el riesgo de un crecimiento sin rendimiento económico satisfactorio", dijo.(CREA, PRENSA: El Norte de Castilla, 2001) '"The economy is, apparently, on the path to recovery very much below the potential rate enhanced by productivity, and even after the measures we took in January, the risk of growth without economic returns remains", he said.' Though relatively less frequently, we also find reportative readings with aparentemente, as in the following: (42) "<...> Ambos expresaron el deseo de que la situación se resolviera con una normalización democrática y hubiera un cese de la violencia The raw and normalized frequencies, specified in Table 1, indicate that supposedly is more frequent in the spoken than in the written subcorpus.On the other hand, the normalized frequency of supuestamente is more than five times higher in the newspaper than in the spoken subcorpus.This important difference is due to its conventionalized use in newspaper discourse to express lack of full commitment as a way of protection against legal problems, especially when dealing with crimes and legal issues, since journalists can easily be made legally responsible for what they write (45).The conventionalization of this use is also registered in Hennemann (2012, 155-157 From the discourse-pragmatic point of view, supposedly and supuestamente often collocate with words that express disappointment, as in ( 46) above, where the speaker hints that the proposition qualified by supposedly could well be true but anyway there was no adequate reaction to the event.Both adverbs are also frequently found in contexts that express scepticism about the validity of the information transmitted: (47) The six adverbs indicating high validity of the information express IIE in all their evidential occurrences except for one of clearly.Not surprisingly, they are often used with the aim of persuading the addressee.By contrast, the adverbs of medium and low validity are more associated with IRE: they commonly mark the information as originating from external voices (report or hearsay), and in so doing they indicate lack of commitment on the part of the speaker/writer.Two exceptions to this tendency are seemingly, which occurs more frequently with IIE values, and aparentemente, for which the number of occurrences of IIE and IRE differ in the spoken and newspaper genres.
In the four subcorpora, the total number of IIE occurrences is higher than that of IRE occurrences except for the Spanish newspaper subcorpus.This distributional difference is largely due to supuestamente, whose conventionalized use in newspaper discourse for evoking external sources of information accounts for the fact that the 50 occurrences in this subcorpus express IRE (see 4.2.3).The predominant type of evidentiality for all the adverbs is the same in the two discourse types, with the exception that aparentemente has a higher number of cases of IIE in the spoken subcorpus and of IRE in the newspaper subcorpus.
The non-evidential cases are slightly more frequent in the spoken than in the newspaper subcorpus.The difference is to be attributed to the higher frequency of evidently and obviously and their Spanish equivalents evidentemente and obviamente for expressing more pragmaticalized and interactional uses rather than an evidential meaning (see Sections 4.1.2and 4.1.3).These facts are congruent with the general higher frequency of these adverbs in spoken discourse (see Table 1).
Non-evidential uses are also frequent in the spoken subcorpus for claramente but for clearly they are more than twice as common in the newspaper subcorpus.This distributional difference is due to the fact that, even though one pragmaticalized case of clearly has been found, namely (11) in Section 4.1.1.,most non-evidential cases have the meaning of manner.As for the high frequency of non-evidential claramente in the spoken subcorpus, it is not due to pragmaticalization but to its frequent use with verbs of communication such as decir ('say'), especificar ('specify'), expresar ('express') or definir ('define').
With regard to the position of the evidential occurrences of the adverbs, specified in Table 3, medial position is the most common in the two English subcorpora and the Spanish newspaper subcorpus, which means that non-focality has had more weight on the whole than iconicity in the choice of position.Conversely, in the Spanish spoken subcorpus, the positions that iconically reflect the meta-operator status of the adverbs, namely initial and parenthetical positions, total 92 cases, while the number of medial occurrences is 51.
Parenthetical position is also strikingly common in the Spanish newspaper subcorpus, where it is used to emphasize or downtone speaker/writer commitment.In contrast to parenthetical position, initial position is more common in the two English subcorpora; therefore, the relative weight of the two positions that highlight iconicity is different in the adverbs in the two languages.As was predicted, the occurrences of final position are scarce, since this position does not highlight either iconicity or information structure: in particular, virtually no occurrences are found in the newspaper subcorpus.In the two languages, inter-adverb differences are also found: al parecer shows by far the highest amount of parenthetical occurrences, followed by apparently, evidently and obviously and their Spanish equivalents, while the other adverbs occur less frequently in this position.This distribution is another instance of the different behaviour of the pair clearly/claramente and the other two pairs of high reliability, which suggests that frequency of parenthetical cases is associated to pragmaticalization.Preference for medial position is strong for the pair supposedly/supuestamente, and also for seemingly.

Conclusions and Suggestions for Further Research
This paper has presented a contrastive analysis of six English evidential adverbs ending in -ly and their Spanish nearest equivalents, based on occurrences in spoken and newspaper discourse from the COCA and CREA corpora.The results of the qualitative and quantitative analyses provide evidence that the adverbs conventionally associated with high reliability, clearly/claramente, evidently/evidentemente and obviously/ obviamente, nearly always express IIE, while the adverbs of medium and low reliability are associated with IRE except for seemingly and aparentemente, which display a more or less balanced number of inferential and reportative occurrences.The number of nonevidential occurrences highly depends on the individual adverb and the contexts of use: they were non-existent for apparently and al parecer, and for supuestamente in the newspaper subcorpus, but quite common with the adverbs of high reliability; these cases are pragmaticized addressee-oriented uses, except for those of clearly and claramente, most of whose non-evidential occurrences have a meaning of manner.The results have also provided evidence that the distribution of clausal positions is quite idiosyncratic for each adverb, but the English adverbs have shown a stronger tendency to occur in medial position while the Spanish adverbs, especially aparentemente, al parecer and evidentemente, have been often found as parentheticals.
The caveat must be made that the number of analyzed cases for each adverb is small, so the results are perforce tentative.Evidentiality has been mainly studied from the perspectives of morphology, syntax and semantics, while its pragmatics is still a relatively unexplored area.More in-depth studies would be needed for each adverb, especially those focusing on their pragmatic values.
Our thanks are extended to two anonymous referees for their thorough reports on a first version of the paper.The remaining flaws and inconsistencies are our sole responsibility.

COCA
To get back, though, to what Bill was saying, I think Bill was onto something very important here in the two videos.The one video clearly <DPE, MD> had them saying that they were now advocates, adherents of Islam.And that would suggest that if you're not an adherent of Islam, therefore you're an enemy, somehow, of Hamas, of the Palestinian people, which is madness.
(2) CONAN: We're talking with David Gergen about the White House and image control.<...> And, David, you mentioned backfiring.Mistakes can happen too --the video released of the soldiers in Iraq apparently <IIE, IN> being prepped for their conversation with the president.Mr-GERGEN: Absolutely, because it becomes too overcontrolled, (COCA, SPOK: NPR_TalkNation, 2005) (c) Indirect-Reportative evidence (IRE): Non-personal indirect access to the evidence provided by other speakers/writers.Reportative markers indicate that the justification for the asserted propositional content derives from information accessed through other speakers/writers.(3) But of course, the official news agencies in Iraq, still controlled, apparently <IRE, PTH>, by Saddam Hussein, are calling this a victory for them.(COCA, NEWS: ABC_Brinkley, 1991) The victim, who was not immediately identified, had apparently <IRE, MD> been renting the small building from the homeowners, DeKalb fire Capt.Eric Jackson.
Kaltenböck (2009) Barron has a soft-spoken, laid-back personality and was dressed in black pants and turtleneck sweater, seemingly relaxed though still in the thick of postproduction editing.(COCA,NEWS:SanFranciscoChronicle,1993)(7) Two weeks ago, in a seemingly typical incident, a correction officer at the James A. Thomas Center was punched by an inmate, "knocked to the floor and rendered unconscious", according to an internal report.(COCA,NEWS:NewYorkTimes,1995)Scopeproperties are also related to position in the clause, asKaltenböck (2009)has observed.The expressions under study have a high degree of positional freedom, since they are found in initial, medial, final and parenthetical positions (cf.Biber et al. 1999,  771; Ruskan 2015, 114-115):-Initial (IN): at the beginning of the clause, in pre-topic constituent position, after conjunctions or other textual elements such as discourse markers:

Table 1 .
Adverbs: Absolute and normalized frequencies (ratio per million words) 2 And my accident occurred one night as I was driving home from work.I was in Harvard Square, and a woman driving a Honda evidently <IIE, MD> wasn't paying attention and smashed into the back of my car, jarring, you know, me out of the driver's seat almost.(COCA, SPOK: NPR, 2014) (21) Massoud has sought to allay such fears, evidently <IIE, IN> preferring to negotiate the fall of Kabul rather than fight a costly battle for it.(COCA, NEWS: Washington Post, 1992) 23) Ray Magliozzi: ez, my first fender bender.Oh, I remember my first fender bender.It was with my dear friend Johnny Mellom (ph), a friend of mine who had a car.And I was -I was in college at the time because in high school I was too young to drive a car -no, no, I was in high school.And he had come to my house in his car, and again it involved a door.I'm bad with doors.Tom Magliozzi: Evidently.
Otherwise though, unless you are given an invitation, don't go --(COCA, SPOK:  CNN_Iraq, 2003)This sense is clearly present in interactional uses of obviously, where the speaker appears to evoke common knowledge or expectations about natural and/or proper behaviour.

Adverbs of Medium and Low Evidentiary Validity 4.2.1 Apparently and Al parecer
): (30) Hasta los seis sí era lo que se podía llamar el principito feliz.Porque obviamente <NE> tenía unos padres que me querían, una hermana mayor con la que jugaba, en fin, un país ideal, un campo precioso, no se podía pedir más, digamos.(CREA, 09.Reportajes: Cartelera, TVE 1, 1996) 'Until the age of six I was what you might call a happy little prince.Because obviously I had parents loved me, an elder sister with whom I played, in fact, an ideal country, a beautiful contryside, you couldn't ask for more, let's say.' The meanings of apparently are defined by the OED as follows: " †1.Evidently or manifestly to the sight; visibly, openly.Obs.<...>; 2. Evidently or manifestly to the understanding; clearly, plainly.<...> 3. To external appearance; seemingly.;<...> 4. So far as it appears from the evidence; so far as one can judge; seemingly."Inferentialevidentialuses of apparently seem to draw on the notions of both observable evidence and judgement, as illustrated in the following example.(31)InWednesday'sshooting, Paul Brandin, 51, of Denver, pulledinto the intersection just after the shooting and got out of his car.He saw a man apparently <IIE, IN> helping another person in a vehicle who had been shot.# He saw another man tussling with the apparent gunman.# (COCA, NEWS: Denver Post, 2006) 34) I just looked into it, it turns out I'm not responsible for it.Justice has been served.(APPLAUSE)JOEL-MCHALE# He's going to kill me.TERRY-GROSS# But apparently <IIE, IN> he didn't kill you, Joel McHale.JOEL-MCHALE# He did not kill me, thank God. That's the first time -I haven't heard that.(COCA,SPOK: NPR, 2015)The meanings of al parecer, according to the DLE (RAE), are the following: "a lo que parece, o al parecer: 1. locs.conjunts.U. para explicar el juicio o dictamen que se forma en una materia, según lo que ella propia muestra o la idea que suscita" (1.conjunctive locutions used to explain the judgement or assessment that is formed regarding some matter, according to what it shows itself or the idea it invokes).It seems that many of them have been taken to some nearby hills, the hills that we have close by, that are called the hills of Chada.These hills, apparently, there are clandestine cemeteries, according to the people in the area, and approximately, well a hundred or a hundred and twenty people whom their relatives know nothing (35) Parece que muchos de ellos han sido llevados a unos cerros próximos, los cerros que tenemos aquí al lado, que son llamados los cerros de Chada.Estos cerros, al parecer <IRE, PTH>, hay cementerios clandestinos según las gentes de la zona, y aproximadamente, pues cien o ciento veinte personas de las que sus familiares no saben nada.(CREA, 09.Reportajes: Informe Semanal, 1986) 'muerte, comprando un cadáver con la cara desfigurada (al parecer <IIE, PTH>, con dinero se puede comprar casi todo), que es identificado como Karol por todos sus familiares y conocidos, que están en el ajo.(CREA, 04.Cine y vídeo: ) # In Texas, Mr. Rove has seemingly <IRE, MD> made few mistakes.Mr. McDonald and other consultants noted that when Mr. Rove set up shop in Austin two decades ago, Democrats held the lion's share of statewide offices.# Now, Republicans do, "and they're all Karl's candidates," Mr. McDonald said, ... (COCA, NEWS: New  York Times, 2000)

3 Supposedly and Supuestamente The
'...Both expressed their wish that the situation should be resolved with democratic normalisation and that there should be an end to violence", Miguel Ángel Cortés said yesterday to this paper, at the time Secretary of State of Coperation with South America in the Government of José María Aznar during the coup d'état against President Hugo Chávez, which apparently resulted in his resignation on 11 April 2002.' meaning of these two adverbs is captured in the definitions provided by the OED: "1.According to belief, but without conclusive evidence and perhaps mistakenly; as is (or was) supposed; by way of supposition.2.As a pretence, by way of feigning.Obs."ThisdefinitionagreeswithCelle's(2009, 287)statement that supposedly is "used when saying what many people say or believe is true, especially when you disagree with them."Supposedly and supuestamente are evidential when the speaker/writer entertains the belief expressed in the proposition, as in (43), and non-evidential when s/he expresses or unmistakably implicates that the proposition is or turns out to be false, as in (44).
). 6 (45) En marzo de 1983 supuestamente <IRE, MD> participó en el secuestro del empresario, diplomático y entonces presidente del Banco de Descuento Diego de Prado y Colón de Carvajal (CREA, NEWS: 09.Justicia, Legislación.El País, 2003) 'In march 1983 [he] supposedly took part in the kidnapping of the businessman, diplomat and then Discount Bank president Diego de Prado y Colón de Carvajal' In the cases of IIE, the context often suggests that the inference is not significant for the flow of discourse or not especially meritorious in terms of mental effort, as in (46), which belongs to the narration of the plot of a film: (46) Lee glances her way, then finally asks her to leave.She goes, supposedly <IIE, MD> to the principal's office, although no one tracks her and to students that notion seems almost comical.The chatty girl returns 15 minutes later and continues to engage classmates from her seat.(COCA, NEWS: Washington Post, 2007)

Table 2 .
Evidential and non-evidential uses of the adverbs in spoken and written discourse (raw numbers and percentages)